Questions tagged [structural-equation-modeling]
Structural Equation Modeling is a multivariate technique. It is based on formulating a set of linear relations between variables, some of which may be latent, and estimating the whole system, typically by analyzing the covariance matrix of the observed variables.
1,196 questions
1
vote
1
answer
27
views
Why are two CFA models not equivalent if I drop some variables from my dataframe or restrict factor loadings to 0, keeping all variables? lavaan
Exploring nested vs. non-nested models, I was surprised with a strange result. The first model uses 21 variables from my dataframe and runs a CFA. The second model uses 28 variables from my dataframe, ...
- 194
0
votes
0
answers
16
views
Negative Lambda in box/cox: Caveats and Interpretation [closed]
I have a highly-skewed DV which has not responded well to standard transformations (square root, cubed root, log or ln, hyperbolic arcsine).I do get some reduction in skew, but not enough to normalize ...
1
vote
1
answer
33
views
Modeling longitidunal relationships without estimating indirect effects
I'm analyzing a three-time point study using structural equation modeling (SEM). I only care about the direct effects of my variables. Is it okay to only model those direct paths, or should I include ...
- 13
0
votes
0
answers
7
views
Prediction with categorical data in semopy [migrated]
I'm using semopy for the first time (I am more familiar with lavaan in R). I was able to apply the ...
- 119
2
votes
2
answers
29
views
Calculating path coefficients manually from correlations
As I understand it, in a path model, we can use Wright's approach to manually calculate (standardized) path coefficients from the correlations between the variables. In the saturated case, we ...
- 41
1
vote
0
answers
34
views
Errors in constructing SEM using lavaan package
I’m looking for a second set of eyes to help diagnose the issues I am having with my SEM model:
Here is my SEM code:
...
- 203
1
vote
0
answers
61
views
Causal Discovery Packages Supporting Mixed Data Types and Prior Knowledge
I am new to the field of Causal Discovery and would like to apply standard algorithms to my dataset, which contains both categorical and continuous variables. I am looking for Python packages that can ...
- 63
1
vote
0
answers
14
views
How many clusters are needed for applying cluster-robust standard errors in a SEM model?
I’m working on an SEM model with data from 1078 students, distributed across 13 schools (clusters). I’d like to apply cluster-robust standard errors, but I’m unsure whether 13 clusters is enough for ...
- 81
1
vote
0
answers
92
views
Best way to report indirect effects in a complex Structural Equation Model (SEM) with a large number of variables and paths
I am fitting a complex structural equation model (SEM) with 4 latent and 2 observed variables (15 direct and 36 indirect effects). Importantly, in this model, some indirect effects from a given ...
- 79
2
votes
0
answers
38
views
Handling Underrepresented Gender Categories in SEM [closed]
I’m including gender as a covariate in my SEM. My data consists of 1,078 cases
with 29 "Prefer not to say" and 8 "Diverse gender".
I’ve treated "Prefer not to say" as NA. ...
- 81
2
votes
1
answer
86
views
Structural Equation Modelling - Semopy
I have created some simulated data. To keep things simple, lets say I have done the following:
y = (X_1 + X_2 + X_3)/3 + e
Where each X_i is drawn from a normal distribution and I have added ...
- 93
0
votes
0
answers
92
views
Using classifier prediction as independent variable with within-subjects data?
In my work I've stumbled upon an interesting result when a classifier is applied to within-subjects data. My question is whether this is a known result, and if so, does it have a name? I can't find ...
- 1,944
7
votes
2
answers
254
views
Can you have bidirectional correlation in path analysis?
Something like this:
Where B causes A, C causes D, and A and D are directly correlated.
Is it possible to solve a system like this?
- 375
1
vote
1
answer
28
views
Path analysis indirect correlation highly correlated to direct one. Is this expected? [closed]
Assume B has a direct effect on A and C has a direct on A as well. Here γ, α and β are the corresponding correlations. β is an indirect "artificial" correlation and γ*α is "true" ...
- 375
3
votes
1
answer
40
views
Is it possible to model a Second-order growth slope factor as a mediator of another second-order growth slope factor?
I am running a second-order growth model with three variables.
My data spans 7 waves of data. I am currently running a multivariate growth model
which models the associative relation of multiple ...
- 31
11
votes
4
answers
647
views
Generally, are we supposed to remove all nonsignificant paths in an SEM analysis?
Given that we favour parsimony in science, should we remove nonsignificant paths that are nonsignificant if they do not result in significantly worse model fit?
After we remove the nonsignificant ...
- 331
5
votes
1
answer
229
views
In SEM, is there any point in interpreting path coefficients if the model shows poor fit in the first place?
If there were some path coefficients that were significant but the whole model shows poor fit, can I still use the result from the significant path?
By extension, if the model evidenced poor fit in ...
- 331
1
vote
1
answer
55
views
How/where to add covariates in SEM mediation model (lavaan)?
Say I have a predictor X, a mediator M, and an outcome Y. I also have a covariate W1 that I assume to have a direct effect on X and M (based on conceptual closeness and empirical findings) but not ...
4
votes
2
answers
156
views
Path Analysis with a variable use d as grouping variable and predictor at the same time
I am conducting path analysis using structural equation modeling (SEM) and have a question regarding the use of a grouping variable in my model. One of my key longitudinal hypotheses is whether the ...
- 41
2
votes
1
answer
28
views
Second-Order Latent Variable in SEM
I am planning to examine the effect of instructional quality on well-being using SEM. Well-being is conceptualized as consisting of five dimensions, with the possibility of calculating a general ...
- 81
2
votes
2
answers
99
views
Dealing with Two-Item Factors in CFA
I am setting up a latent variable model for a CFA in lavaan (R), with plans to use the measurement model in a subsequent SEM analysis. The model includes five correlated factors, each with 3-4 ...
- 81
2
votes
1
answer
36
views
How should you specify a bifactor SEM, where one factor only has one item?
I am trying to respecify an SEM in R using the lavaan package to improve model fit. After inspecting the standardized residuals I noticed that item injnorm2 is formulated in conjunctive (i.e. using ...
- 45
3
votes
1
answer
90
views
All correlations meta-analysis
I am new with MetaSEM. I am trying to fit a simple meta-analysis (I just want to analyse 2 correlations; see Figure).
Here you can find the code that I am using with some simulated data but if there ...
- 73
0
votes
1
answer
36
views
Modeling approaches for conditional probability distribution, applied to Propensity Score estimation for IPW (causal inference)
I'm trying to understand and ideally implement the Inverse Probability Weighting approach to estimate a causal effect. My ressources so far have been Pearl's Primer and the book "What If?".
...
- 225
3
votes
1
answer
24
views
Can I replicate SEM results with only variance-covariance matrix of the summated scores?
I am trying to replicate structural equation modeling results from some research papers using just the variance-covariance matrix.
If the authors estimated the models using observed variables from the ...
- 331
1
vote
1
answer
47
views
Mixed Effects CFA using R and lavaan
I want to conduct a CFA with 3 factors and 9 items each. The data contains cluster of highly varying size. (1 < n < 40, 90 clusters in total while 40 clusters only have one observation). I ...
- 11
0
votes
1
answer
42
views
(R) Second-order latent variable CFA not providing standard errors, one variance is negative
I am going to run an SEM model of one second-order latent variable, parental stress, which is made up by two first-order latent variables. Before I run the SEM model, I am verifying the measurement ...
- 11
0
votes
0
answers
18
views
Adding manifest exogenous variable to SEM leads to error
I am new to SEM, and some help would be very appreciated.
I want to specify a model with one latent endogenous variable, five latent exogenous variables, and a manifest exogenous variable. Something ...
0
votes
2
answers
45
views
Longitudinal model with stable personality trait as predictor
We're planning to run a longitudinal analysis using a stable personality trait to predict the development of a particular skill over time. Originally I thought about running a cross panel model, to ...
3
votes
1
answer
33
views
Invariance test for a longitudinal multi-group growth model. Is a group invariance test for all timepoints required?
I have data from 7 waves for a construct that includes samples from two different countries. As a preliminary step for a growth curve analysis, a longitudinal invariance test has already been ...
- 55
3
votes
2
answers
142
views
Why different results for linear regression in sem() from lavaan and lm()?
thanks for reading this.
I am a student trying to understand more about statistics for social science questions. I am trying to build equivalence in R between the sem() function in {lavvan} and the lm(...
- 31
1
vote
1
answer
162
views
Can Lavaan handle multigroup multilevel SEM?
I am wondering if it is possible to run a multigroup, multilevel SEM in Lavaan.
I can run the analysis just as a multigroup and just as a multilevel, but when I try to specify both cluster and group, ...
- 13
1
vote
1
answer
48
views
Mixed signs of coefficient in mediation - what does that mean in fact?
I have an interpretation problem that I have not been able to solve using any of the books I have found on mediation. I understand many aspects of it, but the model in which the signs are mixed up ...
- 333
1
vote
1
answer
32
views
df (mis)calculation with efa models in lavaan
I've been working on some functionality to wrap EFA scripting in lavaan, and noticed some df's that seemed off. I'm now wondering if I'm goofing/overlooking something, or rather, if something ...
- 5,566
0
votes
0
answers
17
views
Moderation in SEM when Scalar and Metric Noninvariance
I would like to know if it is methodologically sound for me to do a moderation analysis, interacting a latent variable and an observed binary variable (LV1 * X ==>Y) when I have metric and scalar ...
1
vote
1
answer
61
views
Second Level Factor in CFA - how to interpret standardized path coefficients
I have seen such a model in publications many times, which I created on my own data. It is a model of three latent features based on observable variables (likert scales) which also has a superior ...
- 333
1
vote
1
answer
42
views
Fixed weights to 1 in confirmatory factor analysis - how significance (p) value is obtained?
I have a question somewhat related to this thread.
I now understand why the value of 1 should be fixed in the confirmatory analysis to one of paths, and I also understand that various tricks can be ...
- 333
2
votes
1
answer
40
views
Negative covariance matrix in RI CLPM (lavaan)
I am currently struggling to solve the problem of negative covariance regarding the latent variables with respect to the random intercepts in my RI CLPM model, which I calculate via lavaan.
Here is my ...
- 21
1
vote
1
answer
115
views
Lavaan Estimator
I did a three-wave study with repeated measured for each wave.I have two questions:
I am doing a CFA for a configural model (testing temporal invariance for each variable of my model at 3 points in ...
- 11
0
votes
1
answer
45
views
Residual variance in SEM
In structural equation modeling, in a path diagram, the sum of the squares of the estimates (when in standardized units), and the residual variance should sum to 1.
(As in here)
Consider the plot ...
- 777
2
votes
1
answer
36
views
When the direction of causation can plausiblely run in either direction, is there a good stat approach to distinguishing relative strength?
I am interested in an instance where it is plausible that ideology influences economic outcome, and also that economic outcomes influence ideology. Assuming that I have good and relavent measures of ...
- 3,247
2
votes
1
answer
35
views
Queries on structural equation model
Can the outcome / dependent variable be made of only one or two items?
Can I merge different scale items say a 5 point & a 7 point likert into a single latent construct?
Also I have a behavioral ...
0
votes
0
answers
28
views
Should I use raw score or z-score for CFA?
I am running confirmatory factor analyses and my variables have different scales. I wonder whether I should use raw score and report factor loadings of std.all ...
- 61
1
vote
1
answer
20
views
Handling Composite Variables in Latent Class Analysis
I am working with secondary survey data that includes 13 dichotomous variables. According to my theoretical model, I need to group these variables into composite variables. However, I understand that ...
- 11
1
vote
1
answer
65
views
Hypothesis Test Finite Sample Spatial Gaussian Mixture Model
I have $n$ observations of pairs $(x, y)$ and three different models I would like to compare. Model0 is nested within Model1. Model0 is also nested within Model2. I would like to do hypothesis ...
- 425
4
votes
1
answer
91
views
What if path c’ has a negative coefficient, but X - > Y share a moderate to high Spearman correlation?
I am running a mediation model including the variables $X$, $M$, and $Y$.
The Spearman correlation between $X$ and $Y$ is +0.4, while the direct path ...
- 307
0
votes
1
answer
32
views
Measurement invariance across groups and time
I want to run a multi-group longitudinal model. Based on prior invariance testing, I can assume scalar invariance across groups and time. So is it appropriate to have both the same intercepts across ...
- 1
0
votes
0
answers
15
views
multiple Mediated moderation
I am trying to determine if this model represents mediated moderation or moderated mediation and why. I have attached a visual representation of the path diagram for $Y_1=\beta_0 +\beta_{X1}X_1+\beta_{...
- 347
0
votes
0
answers
33
views
Specific indirect effects of a non-recursive SEM in lavaan
The structural equation model I am currently analyzing includes an indirect loop and I was unable to find a straightforward way to calculate specific indirect effects for it. A minimal example of a ...
- 1
4
votes
1
answer
226
views
How to get standardized coefficients of Monte Carlo method for indirect effects in lavaan/semTools?
I've run a path analysis using semTools. I'm interested to test indirect effects. I would like to report the results based on Monte Carlo confidence interval. However, the code ...
- 171